International Journal of Civil Engineering (IJCE) ISSN(P): 2278-9987; ISSN(E): 2278-9995 Vol. 3, Issue 1, Jan 2014, 1-6 Š IASET
ACTIVITY BASED MODELING FOR MODE CHOICE ANALYSIS OF TOWNS-A CASE STUDY IN KERALA SOOSAN GEORGE T1, KRISHNAPRIYA M. G2 & MILIMOL PHILIP3 1
Professor, Department of Civil Engineering, MACE, Kothamangalam, Kerala, India 2
Research Scholar, MACE Kothamangalam, Kerala, India
3
M.Tech Student, Government RIT Kottayam, Kerala, India
ABSTRACT Transportation planning is a science that seeks to study the problems that arrive in providing transportation facilities in an urban, regional or national setting and to prepare a systematic basis for planning such facilities. Trip based approach have been used to formulate model in early times. The behavioral inadequacy of the trip-based approach, and the consequent limitations of the approach in evaluating demand management policies, has led to the emergence of the activity-based approach to demand analysis which views travel as a derived demand; derived from the need to pursue activities. We have collected 150 samples, 50 each from three urban areas in Ernakulam district of Kerala, India by house hold survey. The study required data regarding household characteristics, economic status, and trip characteristics. These variables are used for the model formulation. We have analyzed the data using the software, SPSS and a mode choice model was formulated for each of the three places.
KEYWORDS: Activity Based Approach, Mode Choice, Multinomial Logit Model, SPSS, Travel Behavior INTRODUCTION The need for realistic representations of behaviour in travel demand modelling is well acknowledged in the literature. The result has been an increasing realization in the field that the traditional statistically-oriented trip based modeling approach to travel demand analysis needs to be replaced by a more behaviorally-oriented activity-based modeling approach. Travel demand is an important tool in transportation planning which is generated using socio economic data such as house hold size, automobile availability and employment data. Mode choice analysis is the third step in the conventional four-step transportation forecasting model, following trip generation and trip distribution but before route assignment. Mode choice is an important component of a traveler’s decision regarding trip making. Similar to any decision making process, mode choice also includes the factors like decision maker, alternatives, attributes of the alternatives and decision rules for selecting a particular alternative from the available alternatives. This study is an effort to understand the various factors influencing the mode choice in the study area and then to develop separate mode choice models for the three study areas.
LITERATURE REVIEW The activity-based approach requires time-use survey data for analysis and destination. A time-use survey entails the collection of data regarding all activities (in home and out-of-home) pursued by individuals over the course of a day (or multiple days). Travel constitutes the medium for transporting oneself between spatially dislocated activity participations. The examination of both in-home and out-of-home activities facilitates an understanding of how individuals substitute out-of-home activities for in-home activities (or vice-versa) in response to changing travel conditions. This, in
2
Soosan George T, Krishnapriya M. G & Milimol Philip
turn, translates to an understanding of when trips are generated or suppressed. It is important to note that administrating time-use surveys is similar to administrating household travel surveys, except for collection of in-home as well as out of-home activities. The information elicited from respondents is a little more extensive in time-use surveys compared to travel surveys, but experience suggests that the respondent burden or response rates are not significantly different between time-use and travel surveys. The activity-based approach does require more careful and extensive preparation of data to construct entire "sequences" of activities and travel. On the other hand, such intensive scrutiny of data helps identify data inconsistencies which might go unchecked in the trip-based approach .For example, there might be "gaps" in an individual's travel diary because of non-reporting of several trips, these will be identified during data preparation for activity analysis, but may not be identified in the trip-based approach since it highlights individual trips and not the sequence between trips and activities.
STUDY AREA We have chosen our study area as Perumbavoor, Kothamangalam and Muvattupuzha since Perumbavoor is a satellite town of Kochi city and a municipality in Ernakulam District in the Indian state of Kerala. Kothamangalam is the gateway to high range. The highway NH-85 Ernakulam - Madurai- Rameswaram passes through this town. Muvattupuzha is the junction of three districts namely Ernakulam, Kottayam, Idukki.
THEORETICAL FRAMEWORK The Multinomial logit model is more popular and widely used one due to its computational tractability. However, this model application is cautioned because of IIA property. The proposed model formulation to predict the mode choice in the present study is as follows:
Where, Pin= probability of choosing mode, i‟ by decision maker „n‟ from a choice set, Jn Vin, Vkn = Observed component of disutility for modes‟ i‟ &‟k‟ respectively. Multinomial logistic regression is used to analyse relationships between a dependent variable and independent variables. It compares multiple groups through a combination of logistic regressions. The group comparisons are equivalent to the comparisons for a dummy-coded dependent variable, with the group with the highest numeric score used as the reference group. For each independent variable, there would be two comparisons. Multinomial logistic regression provides a set of coefficients for each of the two comparisons. The coefficients for the reference group are all zeros, similar to the coefficients for the reference group for a dummy-coded variable. Thus, there are three equations, one for each of the groups defined by the dependent variable. The three equations can be used to compute the probability that a subject is a member of each of the three groups. A case is predicted to belong to the group associated with the highest probability.
DATA COLLECTION The study required data regarding household characteristics, economic status, and trip characteristics. Household characteristics include gender and family size. Economic status includes their annual income, floor area, educational qualification, type and number of owned vehicle. Trip characteristics include purpose of trip, origin, destination, mode they used. Mode is categorized into private vehicles, public vehicles and by walk. These variables can be used in the model formulation to develop suitable mode choice models. All these data were obtained by household survey. We have
3
Activity Based Modeling for Mode Choice Analysis of Towns-A Case Study in Kerala
collected 150 samples, 50 each from Muvattupuzha, Perumbavoor and Kothamangalam, three most important towns of Ernaulam district of Kerala, India.
PRELIMINARY ANALYSIS From table 1 it is clear that males prefer to travel by private vehicles compared to females, who mostly depend on public vehicles for their travel needs. Table 1: Effect of Gender on Mode Choice Muvattupuzha Private Public 53.3% 38.3% 38 .3% 52.1%
Gender Male Female
Kothamangalam Private Public 63.4% 28 .6% 41.9% 48 .4%
Perumbavoor Private Public 49.0% 28 .9% 19.5% 52.4%
From table 2 it is clear that there is not much influence of family size on mode choice. Table 2: Effect of Family Size on Mode Choice Muvattupuzha Private Public 2 8% 24% 18 % 34% 31% 24% 23% 18 %
Family Size 3 4 5 6
Kothamangalam Private Public 71.9% 21.1% 68 .7% 29.9% 38.1% 47.6% 39.4% 42.4%
Perumbavoor Private Public 52.9% 29.4% 28 .8 % 50.0% 40.7% 35.6% 72.7% 18 .2%
Table 3 shows that public vehicles are mostly used by employed people where as the unemployed people use private vehicles. Table 3: Effect of Occupation on Mode Choice Occupation Employed Unemployed
Muvattupuzha Private Public 32.9% 53.9% 59.5% 35.3%
Kothamangalam Private Public 20.0% 69.5% 2.6% 10.1%
Perumbavoor Private Public 22.7% 55.7% 54.7% 2 .2%
Table 4 shows that highly qualified people use private vehicles where as students and the less qualified ones depend on public vehicles. Table 4: Effect of Qualification on Mode Choice Qualification Student Sslc Predegree Conventional Professionel Masters
Muvattupuzha Private Public 16.1% 61.3% 33.3% 42.9% 30.0% 60.0% 26.7% 66.7% 40.7% 57.0% 77.4% 19.4%
Kothamangalam Private Public 17.3% 74.7% 73.7% 10.5% 90.9% 9.1% 91.7% 8.3% 65.2% 17.4% 75.0% 25.0%
Perumbavoor Private Public 23.6% 59.7% 45.1% 7. 8% 62.5% 25.0% 54.5% 18 .2% 42.3% 34.6% 21.4% 64.3%
As per table 5 high income people has an affinity towards private vehicles than the low income people. Table 5: Effect of Annual Income on Mode Choice Annual Income < 50,000 50,000-99,999 1,00,000-2,49,999 2,50,000-5,00,000
Muvattupuzha Private Public 29.4% 58.8% 28.6% 53.3% 53.9% 35.7% 58% 34. 8%
Kothamangalam Private Public 39.1% 30.4% 50.0% 44.4% 61.1% 35.4% 73.9% 26.1%
Perumbavoor Private Public 41.0% 34.9% 100.0% 0.0% 43.0% 26.6% 37.0% 50.0%
4
Soosan George T, Krishnapriya M. G & Milimol Philip
From table 6 we can understand that shopping and job trips are by private vehicles, but educational trips are done by public vehicles. Table 6: Effect of Purpose on Mode Choice Purpose Job Education Home Shopping
Muvattupuzha Private Public 63% 33.5% 29.4% 57.4% 46.6% 43.6% 42.3% 34.6%
Kothamangalam Private Public 83.6% 12.7% 17.6% 72.5% 59.5% 31.5% 75.0% 6.2%
Perumbavoor Private Public 53.4% 29.3% 25.0% 60.0% 39.1% 37.4% 22.2% 22.2%
EXPLANATORY VARIABLES The explanatory variables used in the mode choice model development using SPSS are as given below. Table 7: Explanatory Variables Symbol EMP UEP ST SS PR CN PRF INC1
Explanation Employed Unemployed Student Sslc Predegree Conventional Proffessional <10000(income)
Symbol Inc2 Inc3 Inc4 Inc5 Jb Pry Edu Hm
Explanation 10000-50000 50001-99999 100000-249999 250000-500000 Job Prayer Education Home
MODEL DEVELOPMENT The main factors influencing the mode choice for trips of the study area are occupational status, annual income, qualification, purpose etc. The coefficients of the formulated model in the three places are summarized in Table 8. Table 8: Parameter Estimates Variables INTERCEPT EMP UEP ST SS PR CN PRF JB PRY EDU HM SHP INC1 INC2 INC3 INC4 INC5
Muvattupuzha Private Public 51.283 17.768 -2.180 12.503 2.085 11.986 -12.480 -9.96 -16.263 -14.068 -3.077 -0.190 -3.312 -0.404 -15.597 -13.266 -18.992 2.527 -5.781 -1.9320 -20.430 3.253 -19.610 -2.866 -20.004 -3.738 -17.168 -15 -17.428 -13.7 -1.420 -0.263 -2.605 -0.486 0 0
Kothamangalam Private Public 16.008 17.005 18.181 18.272 18.762 18.201 -3.758 -2.062 -1.925 .967 -2.397 -.464 -2.184 -.461 -.701 .804 .573 .149 0 0 .804 .794 .752 .576 0 0 -1.209 .974 -.36 .920 .917 1.094 0 0 0 0
Perumbavoor Private Public 16.099 17.078 1.5 2.263 2.429 1.780 1.407 -.638 .770 -1.826 3.211 1.423 1.664 .180 1.656 .252 -.668 -.984 -2.670 -20.059 -.390 -.610 -.604 -.949 0 0 -17.914 -17.095 -.533 -17.880 -18.046 -17.092 -17.644 -16.995 0 0
Activity Based Modeling for Mode Choice Analysis of Towns-A Case Study in Kerala
5
CONCLUSIONS In all the three locations employed persons prefer public vehicle than private, whereas unemployed people have a positive effect on public vehicle. Professionals prefer private vehicle compared to public vehicle in Kothamangalam and Perumbavoor. As far as purpose is considered, people prefer public vehicle for education in the three locations. In case of income, as the income increases the preference to private vehicle increases. The travel characteristics like origin, destination and duration are important but it does not appear in the mode choice models. No such variables are significant in the study. It may be due to the selected study area. Weight ages for the attributes may not suit the socioeconomic characters of the location.
ACKNOWLEDGEMENTS First author thanks Mr. G.Ananthu Krishnan, Miss. Bhagyashree.R, Miss. Leya Asok, Miss. Nethi.R, Miss. Silpa.P.sasikumar and Mr. Vivek.M.S, former B.Tech. Students, Department of Civil Engineering, Mar Athanasius College of Engineering Kothamangalam for their active participation in conducting the studies and preparation of results.
REFERENCES 1.
Chandra R & Frank. S, Activity-based modeling of travel demand, Handbook of Transportation Science, Vol. 1, 1-27, 2006.
2.
Jiangyong.L et al., Modelling on mode choice behavior based on trip chaining: a case study in zhongshan city, icctp 2011 Š asce 2011, Vol 31- 825-835, 2011.
3.
Frank. S & Bhat, A Self Instructing Course in Mode Choice Modeling:Multinomial and Nested Logit Models, Vol 1, 1-249, 2007.
4.
Goran Jovicic, Activity based travel demand modelling- a literature study, Handbook of Transportation Science,, Vol 1,1-64, 2009.
5.
Multinomial logistic regression, SW388R7 Data Analysis & Computers II (ppt).
6.
Philip. A, Modern mode choice analysis, Handbook of Transportation Science, Vol -11 20-45, May 9, 2012.
7.
Lee Sam Fuk Man. A model- based study of the transport planning for new town development in Hong Kong1996 Transportation Planning and Technology Vol-1, 1-127, 1996.
8.
Vovsha Peter et al. , New Advancements in Activity-Based Models, Australasian Transport Research Forum 2011 Proceedings, Vol -11, 1-20, 2011.